BRIF.

Get to know your guest before the conversation.

Get a free brief
Demo example built from publicly available sources. The person is not affiliated with BRIF and does not endorse it.
Imagine this preparation for your next guest.Get your first brief free
YOUR BRIEF · PREPARED 18 OCTOBER 2026 · DEEP BRIF · 50 SOURCES

Sam Altman

CEO of OpenAI, creator of ChatGPT, former Y Combinator president
PODCAST PREP

In 60 seconds

Why this guest

Sam Altman has run OpenAI since 2019; ChatGPT hit 1.2 billion weekly users at DevDay on September 29, 2026.

First question

You have said that the CEO is far from the most important person at OpenAI, and that a handful of researchers are individually more critical than you are. Who inside the company did something in the last month that you are still thinking about?

Don't open with

Take us back to the five days in November 2023 when you were fired and reinstated.

Careful

ChatGPT weekly user count and OpenAI revenue run-rate change frequently — confirm the latest figure on air (1.2 billion weekly users as of September 29, 2026, per DevDay).

The numbers

50sources
23min listened
11questions with a reason
5to confirm live

In this brief

  1. 01Before recording
  2. 02Run of show
  3. 03Already asked
  4. 04The person and the company
  5. 05Before and after the episode
  6. 06Sources

How to read: numbers in brackets [n] mark the source and are a click away. ▶ opens the recording at that minute. All sources are numbered at the end.

01

Before recording

The 60-second briefing

  • Sam Altman has run OpenAI since 2019; ChatGPT hit 1.2 billion weekly users at DevDay on September 29, 2026.[23][24]
  • OpenAI's annualized revenue run rate is about $50 billion as of October 2026, with talks underway for a $1.4 trillion valuation round.[33][32]
  • He says an IPO is 'ill-timed' before 2027, citing safety concerns.[25]
  • Key for the conversation: he presents himself as an evangelist and communicator, not the key technical driver — push past that framing for concrete decisions he personally made.[3]
  • Careful: reports of his 'thaw' with Elon Musk and his switch to polyphasic sleep rest on a single X post each, relayed secondhand — don't treat them as confirmed without asking him directly.[4]

Stage card: the must-ask questions in order

★ = must-ask; the rest add depth if time allows.

  1. You have said that the CEO is far from the most important person at OpenAI, and that a handful of researchers are individually more critical than you are. Who inside the company did something in the last month that you are still thinking about?[3]

  2. After GPT-5 shipped you said OpenAI learned a lesson about upgrading a product for hundreds of millions of people in one day, and 4o access came back within about a day. What is different in the mechanics of a launch now compared with that one?[47]

  3. You have said publicly that if you had double the compute you would be at double the revenue right now. With the compute you actually have, how do you decide what goes to training runs, what goes to free users and what goes to the API?[48][35]

  4. OpenAI's gross margin recovered to about 39% in Q1 2026 from 33% in 2025, against an internal target of 52% by year-end. How do you judge whether a product line is allowed to keep losing money, and when it has to stop?[37][39]

  5. ChatGPT Ads passed a $1 billion annualized run rate within roughly 200 days of launch. What did you personally refuse to let advertising do inside ChatGPT?[37]

  6. OpenAI had about 770 people in November 2023, and outside estimates of your headcount now run anywhere from 4,500 to nearly 10,000. What broke first at that pace, and what did you change because of it?[46][37]

  7. You said OpenAI will definitely build a humanoid robot, and the company is hiring across motion control, actuators and manufacturing. What has to be true before you let a new line like that pull people off the core product?[26][36]

  8. You have said the world should accept some bad things happening for the benefits of this technology. Who at OpenAI decides which bad things are acceptable?[49][8]

  9. You have said 95% of a CEO's job is communication, evangelism and telling the story of the company. For a founder with five people and no audience, what is the one thing you would have them do next week?[3]

Who they are

Sam Altman has run OpenAI since 2019, turning a research lab into the company behind ChatGPT, now used by 1.2 billion people weekly as of September 2026. He dropped out of Stanford to build Loopt, then spent five years leading Y Combinator before returning to AI full time. He is as much a public evangelist for AI as its builder, and draws both fascination and deep skepticism about his honesty and motives.[15][19][24]

Right now

At OpenAI's DevDay on September 29, 2026, Altman announced ChatGPT reached 1.2 billion weekly users and discussed AI safety and the company's path to a possible IPO.[23][24]

Verify live

ChatGPT weekly user count and OpenAI revenue run-rate change frequently — confirm the latest figure on air (1.2 billion weekly users as of September 29, 2026, per DevDay).

Whether the Altman-Musk 'thaw' after the Roadster post is substantive or a one-off gesture is unconfirmed.

Whether polyphasic sleep claim is Altman's own verified statement or media paraphrase should be confirmed directly.

AI safety and existential-risk debates are contentious in his field; treat claims about autonomous AI agents or 'rogue' behavior as disputed, not settled.[2]

Altman's personal honesty has been publicly questioned by former colleagues; distinguish allegation from established fact when raising it.[6]

Producer notes

  • He answers abstract AI questions in polished, rehearsed paragraphs and gets concrete only when you ask for a number, a name or a specific decision — default to 'which one' and 'who decided' over 'how do you think about'.
  • When he says 'we decided' or 'we learned a lesson', push for who was in the room and which argument lost; his GPT-5 answer in particular is a finished sentence that needs a second push to become a process.
  • He is comfortable conceding mistakes in general terms (the 4o rollback) but will not volunteer the specific internal example — ask for the example by name and wait through the pause.
  • Keep safety and AGI philosophy to the two contrarian slots; this audience of founders and operators gets more from pricing, compute allocation, margins and hiring.
  • He is comfortable being the public face and will happily talk about evangelism and communication — use that as the landing for the closing, since it is both his genuine thesis and the most portable advice for a five-person team.
02

Run of show

Timing

0–6opener: the people inside OpenAI he is still thinking about; sets a concrete, non-stump tone.
6–16warm-up: the GPT-5 launch mechanics and the Go/Plus/Pro pricing ladder.
16–45core: compute allocation, margin discipline, the advertising line, scaling from 770 people to today's headcount, the test for starting a new bet.
45–55contrarian: who decides which bad outcomes are acceptable, and what changed with the board.
55–60closing: the one communication move a five-person team should make next week.

Run of show

★ = must-ask; the rest add depth if time allows.

Opener

  1. You have said that the CEO is far from the most important person at OpenAI, and that a handful of researchers are individually more critical than you are. Who inside the company did something in the last month that you are still thinking about?[3]

    Why: Warm, specific, and immediately gets him off the stump speech and into concrete people and work rather than AGI abstractions.

Warm-up

  1. After GPT-5 shipped you said OpenAI learned a lesson about upgrading a product for hundreds of millions of people in one day, and 4o access came back within about a day. What is different in the mechanics of a launch now compared with that one?[47]

    Why: He has given the lesson as a sentence; founders need the process change it produced — rollback plans, staged rollouts, who can pull the trigger.

  2. ChatGPT Go went global in January 2026 at $8 a month, alongside Free, Plus at $20 and Pro tiers. What convinced you that an $8 tier was worth the revenue you give up from people who would have paid $20?[42]

    Why: Pricing-ladder decisions are directly transferable to the founders and operators listening, and this one has a visible cannibalisation risk.

Core

  1. You have said publicly that if you had double the compute you would be at double the revenue right now. With the compute you actually have, how do you decide what goes to training runs, what goes to free users and what goes to the API?[48][35]

    Why: Every founder rations a scarce resource; this makes him show the allocation rule rather than repeat the constraint.

  2. OpenAI's gross margin recovered to about 39% in Q1 2026 from 33% in 2025, against an internal target of 52% by year-end. How do you judge whether a product line is allowed to keep losing money, and when it has to stop?[37][39]

    Why: Operators want the stopping rule, not the growth story; this is the most applicable number in the whole dossier.

  3. ChatGPT Ads passed a $1 billion annualized run rate within roughly 200 days of launch. What did you personally refuse to let advertising do inside ChatGPT?[37]

    Why: Creators and operators care where the line sits between monetisation and product trust, and he is the one who sets it.

  4. OpenAI had about 770 people in November 2023, and outside estimates of your headcount now run anywhere from 4,500 to nearly 10,000. What broke first at that pace, and what did you change because of it?[46][37]

    Why: Scaling pain is the most portable lesson he has for a founder audience, and he has lived three orders of magnitude of it.

  5. You said OpenAI will definitely build a humanoid robot, and the company is hiring across motion control, actuators and manufacturing. What has to be true before you let a new line like that pull people off the core product?[26][36]

    Why: Gives a concrete test for when a company is allowed to start a second bet — the question every scaling founder is arguing about internally.

Contrarian

  1. You have said the world should accept some bad things happening for the benefits of this technology. Who at OpenAI decides which bad things are acceptable?[49][8]

    Why: His own formulation invites the obvious unanswered question — the decision procedure behind it, not the philosophy.

  2. In the New Yorker profile by Ronan Farrow and Andrew Marantz, a former board member described you as unconstrained by truth. What have you changed in how you communicate with your board since that period?[6]

    Why: The allegation is public and central to how founders read him; asking about the process change is fair and answerable without presuming the charge is true.

Closing

  1. You have said 95% of a CEO's job is communication, evangelism and telling the story of the company. For a founder with five people and no audience, what is the one thing you would have them do next week?[3]

    Why: Turns his best-known thesis into a concrete action for the people listening, which is the takeaway the show wants.

Don't ask — and what to ask instead

  • Take us back to the five days in November 2023 when you were fired and reinstated.Why not: He has told this story in nearly every major interview and will give the settled version, burning ten minutes.Instead: What do you check now, in your board relationship, that you did not check before 2023?
  • What is going on between you and Elon Musk?Why not: Active litigation and a feud he answers with the same lines; it produces a headline, not a lesson for founders.Instead: What did the restructuring into a public benefit corporation change about who can overrule you?
  • Is it true you switched to polyphasic sleep because GPT-5.5 in Codex is so good?Why not: This circulates through aggregator coverage and is explicitly flagged as unverified; building a question on it risks a correction on air.Instead: What does a normal working week look like for you now compared with your Y Combinator years?
  • When will AGI arrive, and how do you define it?Why not: Answered at TED, on Hard Fork, with Lex Fridman and with Tucker Carlson; the audience has heard every version.Instead: What capability would make you personally delay a release by a quarter?
  • When is the IPO?Why not: He has already ruled out 2026 and said 2027 at the earliest; the answer is on the record and he will not move off it.Instead: What does OpenAI have to be able to report monthly before it could be a public company?
  • Do you think you are becoming too powerful, or playing God?Why not: Frames the conversation as an accusation, puts him into defensive talking points and teaches founders nothing.Instead: Who inside OpenAI can tell you no, and when did that last happen?
  • Tell us about your 2023 world tour of 28 countries.Why not: He has already been walked through the decision and the logistics, down to whether the plane had a bed.Instead: What did meeting that many heads of state change about how you talk to regulators now?

Likely questions from the audience

  1. Should a small startup build on top of OpenAI when you might ship the same feature?

    He frames infrastructure, not demand, as the binding constraint and points to Codex at 3 million weekly users and 15 billion tokens per minute through the API — steer him to where he thinks the API layer leaves room.

  2. What does a founder do when their product's unit economics do not work yet?

    Gross margin was 33% in 2025 and about 39% in Q1 2026, with a -122% Q1 2026 operating margin, while raising at $852 billion — press for his actual threshold for tolerating losses versus stopping.

  3. How do you price a product when nobody knows what it is worth?

    Free, Go at $8, Plus at $20 and Pro up to $500, plus localized pricing such as ₹399 in India — ask what evidence moved each price.

  4. What is the first hire a founder should make once the product is working?

    He has argued every startup needs at least one evangelical founder, usually the CEO, who recruits, sells, talks to press and raises money — ask what he hires around that person.

  5. Will there still be room for human creators when models make video and music?

    An AI film associated with him drew enough controversy that Amazon dropped it — a live, concrete hook rather than an abstract debate.

Backup questions

  • You co-founded Merge Labs, which came out of stealth in January 2026 with a roughly $252 million seed led by OpenAI. How do you decide how many hours a week that is allowed to take?
  • Enterprise is now more than 40% of OpenAI's revenue and is tracking toward parity with consumer. What had to change in how you sell for that to happen?
  • OpenAI partnered with Deutsche Telekom to reach 261 million customers, after deals with T-Mobile, Orange and SK Telecom. What do telecoms ask for in those negotiations that you did not expect?
03

Already asked

Most-asked questions

Listened to: 1/3 · 23 min

  • Is Sam Altman trustworthy / 'unconstrained by truth'? ×1Decoder with Nilay Patel

Fresh angles — ask these

  • You have said publicly that if you had double the compute you would be at double the revenue right now. With the compute you actually have, how do you decide what goes to training runs, what goes to free users and what goes to the API?[48][35]
  • OpenAI's gross margin recovered to about 39% in Q1 2026 from 33% in 2025, against an internal target of 52% by year-end. How do you judge whether a product line is allowed to keep losing money, and when it has to stop?[37][39]
  • ChatGPT Ads passed a $1 billion annualized run rate within roughly 200 days of launch. What did you personally refuse to let advertising do inside ChatGPT?[37]
  • You said OpenAI will definitely build a humanoid robot, and the company is hiring across motion control, actuators and manufacturing. What has to be true before you let a new line like that pull people off the core product?[26][36]
  • You have said the world should accept some bad things happening for the benefits of this technology. Who at OpenAI decides which bad things are acceptable?[49][8]

Topics they light up on

  • CEO as evangelist, not the most critical person at OpenAI – States plainly that researchers matter more than he does to OpenAI's success, a recurring line across his YC-era talks.[3]“the CEO is far from the most important person in the company... there's a large handful of researchers, each of which are individually more critical to the success we've had so far”

Topics they handle carefully

  • Whether his 2023 world tour reflected a personal strategy – Gives brief, logistical answers ('I had a bed') rather than elaborating on intent, per Nudge's archival clip.

Unexplored stories

  • Altman's reported switch to polyphasic sleep to keep pace with Codex/GPT-5.5[4]mentioned only via a secondhand X-post summary in Biography Flash, unverified directly
  • Apparent thaw with Elon Musk after the Roadster dispute[4]a single conciliatory X post reported in Biography Flash, with no elaboration from Altman

What they have already said publicly

  • Says 95% of a CEO's job is communication and evangelism, not day-to-day execution.[3]
  • Describes himself as far from the most important person at OpenAI, crediting researchers instead.[3]
  • Recounts a 2023 world tour of 28 countries in 35 days, meeting heads of state including Modi and dining with Rishi Sunak.[3]
  • Admitted OpenAI 'totally screwed up' the abrupt GPT-4o-to-GPT-5 transition, restoring GPT-4o access within about a day after user backlash.[47]
  • Would stand up to his own investors rather than let unsafe AI development proceed.[25]
  • Argues the world should accept some negative consequences of AI for the sake of its benefits and people's agency.[49]
  • Says infrastructure, not demand, is OpenAI's binding constraint, claiming doubling compute would double current revenue.[48]
  • Confirmed OpenAI will build a humanoid robot, saying the form factor makes sense because the world is built for people.[26]

In their own words

“Another thing that startups need at least one evangelical founder usually the CEO and this requires someone who can infect with enthusiasm the whole world about what the company is trying to do.”

Do AI bosses need to be famous?

“To be an effective CEO, like 95% of your job is effective communication and evangelism and telling the story of the company.”

Do AI bosses need to be famous?

“The CEO is far from the most important person in the company... there's a large handful of researchers, each of which are individually more critical to the success we've had so far.”

Do AI bosses need to be famous?

“i am switching to polyphasic sleep because GPT-5.5 in codex is so good that i can't afford to be sleeping for such long stretches and miss out on working”

↗Biography Flash: Sam Altman on AI Rivals Lawsuits and the Road to AGI
04

The person and the company

Story

  • Enrolled at Stanford to study computer science in 2003, dropped out in 2005 after Y Combinator accepted his pitch for its Summer Founders Program.[18]
  • Co-founded location-sharing startup Loopt in 2005 and ran it as CEO until it was acquired by Green Dot for $43.4 million in March 2012.[19][15]
  • Co-founded venture fund Hydrazine Capital with his brother in 2012.[18]
  • Joined Y Combinator as a part-time partner in 2011, became president in February 2014 succeeding Paul Graham, and led it until March 2019, funding close to 1,900 companies.[19][20]
  • Co-founded OpenAI in December 2015 as a nonprofit with Elon Musk, Greg Brockman, Ilya Sutskever and others.[27][28]
  • Became CEO of OpenAI in 2019; was removed by the board in November 2023 and reinstated five days later.[15]
  • Chaired Oklo's board from 2015 and served as its CEO from July 2021 to May 2024, stepping down as chairman in April 2025.[21][22]
  • Chaired Helion Energy's board for over a decade before stepping down in March 2026 amid talks for Helion to sell power to OpenAI.[20]
  • Co-founded Merge Labs, a brain-computer interface startup, which came out of stealth in January 2026 with a $252 million seed round led by OpenAI.[16][17]
  • OpenAI completed a recapitalization on October 28, 2025, converting its for-profit arm into a public benefit corporation controlled by a nonprofit foundation holding about 26% ownership.[29][28]

Interests & hobbies

AGI and superintelligenceAI safety and risknuclear energybrain-computer interfacesstartups and evangelismrobotics

Recent (the year before this brief)

  • At DevDay on September 29, 2026, announced ChatGPT has 1.2 billion weekly users, up from 900 million in February 2026.[24]
  • Said on September 12, 2026 that an OpenAI IPO is 'ill-timed' given safety concerns and would not happen before 2027, and that he would stand up to investors if needed.[25]
  • Confirmed in early September 2026 that OpenAI will build a humanoid robot: 'We will definitely do a humanoid.'[26]
  • Posted on X on October 16, 2026 calling Tesla's new Roadster 'quite amazing,' a conciliatory gesture after a public dispute with Elon Musk over a cancelled Roadster reservation.[4]
  • Reportedly said in October 2026 he was switching to polyphasic sleep because GPT-5.5 in Codex is too capable to let him sleep through long stretches; this is a secondhand report that should be confirmed directly.[4]

Company and context

  • OpenAI was founded in December 2015 as a nonprofit by Altman, Elon Musk, Greg Brockman, Ilya Sutskever and others, with a mission to ensure AGI benefits humanity.[27][28]
  • OpenAI completed a recapitalization on October 28, 2025, becoming a public benefit corporation majority-influenced by a nonprofit foundation holding about 26% ownership.[29][28]
  • OpenAI closed a $122 billion funding round on March 31, 2026 at an $852 billion valuation, led by SoftBank with Amazon, Nvidia and Microsoft participating.[30][31]
  • Headcount estimates vary widely: the March 2026 plan was to grow from 4,500 to 8,000 by year-end 2026, while Revelio Labs put the workforce at about 7,850 in December 2025 and a LinkedIn-based tracker at 9,837 in November 2025.[28][46]
  • ChatGPT is currently the fifth-most-visited website globally, as of September 2026.[28]
  • OpenAI is reportedly in talks for a new funding round of at least $30 billion at a roughly $1.4 trillion valuation as of October 2026.[32][34]
  • OpenAI's Microsoft partnership was restructured on April 27, 2026 into a non-exclusive deal, removing Microsoft's uncapped revenue share while capping OpenAI's reverse share through 2030.[40]

Company timeline

OpenAI's official annualized revenue was $2 billion in 2023.
ChatGPT weekly active users passed 200 million by August 2024.
OpenAI's annualized revenue reached $6 billion in 2024 (official ARR figure), while recognized full-year 2024 revenue was $3.7 billion.
ChatGPT weekly active users reached 400 million by February 2025.
OpenAI closed 2025 with annualized revenue of more than $20 billion (OpenAI official) or $21.4 billion per The Information, on leaked audited full-year 2025 revenue of $13.07 billion and an operating loss of roughly $20.9 billion.
On April 27, 2026, Microsoft and OpenAI announced a restructured, non-exclusive partnership: Microsoft stopped paying OpenAI a revenue share, while OpenAI's reverse revenue share to Microsoft continues through 2030 but is now capped rather than uncapped.
OpenAI closed a $122 billion funding round on March 31, 2026 at an $852 billion post-money valuation, anchored by Amazon, Nvidia and SoftBank.
OpenAI's annualized revenue run rate reached $25 billion by February 2026, up 17% from $21.4 billion at the end of 2025.
ChatGPT weekly active users reached 900 million by February 27, 2026.
One financial tracker put the annualized revenue run rate near $70 billion in late September 2026, while CNBC reported about $50 billion in October 2026 — treat the figure as moving. Headcount estimates for the same period ranged from 4,500 to nearly 10,000 against a target of 8,000 by year-end.

How the company grows (public — don't ask again)

  • Pricing ladder: ChatGPT is Free ($0), Go ($8/month), Plus ($20/month) and Pro (from $100 to $500/month depending on tier), with Go localized in some markets such as ₹399 per month in India.[42]
  • ChatGPT Go launched in India in August 2025 at roughly $4-5/month, then rolled out globally on January 16, 2026 at $8/month and is now available in about 98 countries with localized pricing in some markets.[42]
  • App Store pricing varies widely by country, from PHP 999 (about $15.93) in the Philippines up to HUF 8,990 (about $28.46) in Hungary, a roughly 79% spread, as checked across 39 markets on September 16, 2026.[43]
  • Distribution via telecoms: OpenAI struck a multi-year strategic partnership with Deutsche Telekom (announced around September 8, 2026) to deploy AI to 261 million customers, with pilots in Q1 2026, following earlier telecom deals with T-Mobile, Orange and SK Telecom.[44]
  • Enterprise channel partnerships: IBM announced a strategic partnership with OpenAI on August 13, 2026 that includes joint go-to-market initiatives and a dedicated OpenAI Practice with thousands of IBM consultants trained on OpenAI Partner Network certifications.[45]
  • Advertising as a new revenue line: OpenAI's ChatGPT Ads surpassed a $1 billion annualized run-rate within about 200 days of launch, as of August 31, 2026.[37]
  • Funding stance: rather than bootstrapping, OpenAI raised $122 billion in its March 2026 round at an $852 billion valuation and was in talks by late September 2026 to raise at least $30 billion more at a roughly $1.4 trillion pre-money valuation, with a five-year plan projecting $278 billion of cumulative negative free cash flow through 2030.[37]
  • Team scale: OpenAI grew from about 770 employees in November 2023 to several thousand today — estimates range from 4,500 to nearly 10,000 depending on the tracker, against a stated target of 8,000 by the end of 2026 — with hiring spread across research, engineering, product, legal, policy and global operations.[46]
  • Unit economics: company-wide gross margin fell to 33% in 2025 from 40% in 2024, missing an internal 46% target, then recovered to about 39% in Q1 2026 against a new target of 52% by year-end 2026; Q1 2026 operating margin was -122%, losing about $1.22 for every dollar earned.[37]
  • What did not work: Altman admitted the abrupt discontinuation of GPT-4o in favor of GPT-5 alienated users who had formed emotional attachments to the older model, forcing OpenAI to restore GPT-4o access within about a day.[47]
  • Infrastructure financing is largely partner-funded rather than from OpenAI's own balance sheet, including Oracle's five-year Stargate compute commitments, while OpenAI itself has committed up to $250 billion in Azure spend.[37]

Public positions — build on them, don't repeat

  • Altman argues infrastructure, not demand, is the binding constraint on OpenAI's growth, saying doubling compute would double current revenue. — “I think if we had double the compute, we'd be at double the revenue right now”[48]
  • Altman states he would be willing to defy his own investors rather than let unsafe AI development proceed.[25]
  • Altman argues some negative consequences of AI are an acceptable tradeoff for the technology's benefits and for giving people agency. — “world should accept some bad things happening for the benefits of this technology and people having the agency”[49]
  • Altman frames the current moment as an inflection point he describes using the word 'singularity,' tying it to AI increasingly contributing to AI research itself. — “We are now, like, in the singularity”[50]
  • Altman says deploying a single model change to hundreds of millions of users at once carries risks OpenAI underestimated with GPT-5, shaping a more cautious rollout approach going forward. — “We've learned a lesson about what it means to upgrade a product for hundreds of millions of people in one day”[47]

What's in it for the guest

  • He is actively promoting OpenAI's robotics push and the 1.2 billion weekly user milestone — a natural hook that lets him talk about product traction without sounding promotional.[23][24]
  • He is also navigating public scrutiny over AI safety and IPO timing, having said in September 2026 that an IPO is 'ill-timed' — a founder-relevant angle on weighing growth pressure against caution.[25]
  • Frame him as the operator-evangelist he describes himself as ('95% of your job is communication') rather than purely a researcher or technologist.[3]

Ways in nobody uses

  • He has said the CEO is 'far from the most important person' at OpenAI — worth probing concretely: what does he do day to day if not research?[3]
  • His own line that doubling compute would double revenue suggests a quantified view of the company's bottleneck — a rare concrete growth claim to press on.[48]
  • The GPT-4o rollback is a specific, admitted operational mistake — an unusually candid 'what I'd do differently' moment to build on.[47]

Appearances & sources — detailed

Three per-source analyses were read in depth (one archival-clip compilation, one AI-aggregated news roundup with no direct Altman speech, and one episode where Altman was discussed by a journalist rather than interviewed); the remaining listed appearances were not analysed in detail, so this brief leans heavily on the dated dossier for current facts.

05

Before and after the episode

Invitation to the guest

A draft that proves you did the homework. Adapt and send.

ChatGPT now reaches 1.2 billion weekly users, and I'd like to build the conversation around decisions rather than predictions. We won't ask again whether Sam is trustworthy or whether AGI is near — those get asked every week. Instead we want to dig into how he runs OpenAI day to day, what the GPT-4o rollback taught him about shipping to hundreds of millions of people, and where the compute bottleneck really bites. The show reaches founders, operators and builders who want methods, not predictions. Let me know what date might work.

After the episode

Thank-you note (within a day)

Thanks for making time to record — [specific moment, e.g. the GPT-4o rollback story] was the kind of concrete detail our audience rarely gets. We'll let you know the exact air date once it's locked, and send the final cut for your review beforehand.

What you can give them

  • Send a clip of his strongest concrete answer (e.g. on the GPT-4o rollback) for his own channels.
  • Offer a shout-out of OpenAI's current robotics hiring push if relevant to listeners.
  • Flag any founder in the audience building in BCI or nuclear energy who might be a useful OpenAI/Merge Labs/Oklo connection.

When to get back in touch, and why

Around early 2027, when OpenAI's humanoid robot demo is expected, follow up to ask how that build went.

06

Sources

  1. [1]The Bumpy Road To AGI — With Kevin Roose ↗podcasts.apple.com
  2. [2]AI Insider: "They Are Hiding What Is Really Happening" | Jeffrey Ladish ↗podcasts.apple.com
  3. [3]Do AI bosses need to be famous? ↗podcasts.apple.com
  4. [4]Biography Flash: Sam Altman on AI Rivals Lawsuits and the Road to AGI ↗podcasts.apple.com
  5. [5]OpenAI Bugging Incident, Are Your Devices Listening? Xbox Launches Remix Lab | Diet TBPN ↗podcasts.apple.com
  6. [6]Rewind: Ronan Farrow on Sam Altman’s ‘unconstrained’ relationship with the truth ↗podcasts.apple.com
  7. [7]The Dark Side of Silicon Valley: AI Gods, Sex Bots, and Propaganda feat. Bridget Todd ↗podcasts.apple.com
  8. [8]Why Altman Says "Accept Some Bad Things Happening" and What It Means for AI Safety | MOONSHOTS #301 ↗podcasts.apple.com
  9. [9]🐟 ”Tuna Tuesday” — Sushi Carving surge. Sam Altman’s AI movie. Red-Dye Diesel’s drama. +Warren Buffett YouTube ↗podcasts.apple.com
  10. [10]OpenAI's Sam Altman Talks ChatGPT, AI Agents and Superintelligence — Live at TED2025 ↗youtube.com
  11. [11]Sam Altman Shows Me GPT 5... And What's Next ↗youtube.com
  12. [12]Sam Altman on God, Elon Musk and the Mysterious Death of His Former Employee ↗youtube.com
  13. [13]'Hard Fork': An Interview With Sam Altman ↗youtube.com
  14. [14]Sam Altman: OpenAI CEO on GPT-4, ChatGPT, and the Future of AI (Lex Fridman Podcast) ↗youtube.com
  15. [15]Sam Altman - Wikipedia ↗en.wikipedia.org
  16. [16]OpenAI Leads $252 Million Bet on Merge Labs Brain-Computer Interfaces - Hypebeast ↗hypebeast.com
  17. [17]OpenAI backs Sam Altman's brain computer interface startup Merge Labs - BusinessToday ↗businesstoday.in
  18. [18]Sam Altman - Britannica ↗britannica.com
  19. [19]Sam Altman: OpenAI CEO Sam Altman Biography, Net Worth, Education and more - Interesting Engineering ↗interestingengineering.com
  20. [20]Who Is Sam Altman? OpenAI CEO Career and Investments - Layer3Labs ↗layer3labs.io
  21. [21]Oklo Inc. - Form S-1/A - FY2024 (SEC) ↗sec.gov
  22. [22]Oklo Inc. - Form 424B3 - FY2024 (SEC) ↗sec.gov
  23. [23]Watch CNBC's full interview with OpenAI CEO Sam Altman (CNBC) ↗cnbc.com
  24. [24]ChatGPT Statistics Q3 2026: 1.2 Billion Weekly Users (TechnologyChecker.io) ↗technologychecker.io
  25. [25]Exclusive: Sam Altman addresses AI doomsday fears, IPO timing in new Fortune interview ↗fortune.com
  26. [26]OpenAI is building a humanoid robot, but the body isn't the real challenge (Engadget) ↗engadget.com
  27. [27]Sam Altman Biography - biyografiler.com ↗biyografiler.com
  28. [28]OpenAI - Wikipedia ↗en.wikipedia.org
  29. [29]Our Structure - OpenAI ↗openai.com
  30. [30]OpenAI Revenue and Valuation Statistics (October 2026) ↗explodingtopics.com
  31. [31]OpenAI funding: Every round, investor, and valuation (2026) ↗revenuememo.com
  32. [32]OpenAI reportedly in talks to raise $30B round at $1.4T valuation ↗techcrunch.com
  33. [33]Nvidia, Oracle, CoreWeave and other AI stocks sink on OpenAI revenue report (CNBC) ↗cnbc.com
  34. [34]OpenAI Expects $70 Billion in Annualized Revenue by End of 2026 (Yahoo Finance/Bloomberg) ↗finance.yahoo.com
  35. [35]The next phase of enterprise AI ↗openai.com
  36. [36]Altman confirms OpenAI will build humanoid, demo in 2027 (aiweekly.co) ↗aiweekly.co
  37. [37]OpenAI Revenue 2026: ~$70B ARR, Q2 Loss & $1.4T Valuation ↗useluminix.com
  38. [38]ChatGPT Statistics 2026 - Superlines ↗superlines.io
  39. [39]OpenAI Revenue 2026: $25B ARR and a -122% Operating Margin ↗valueaddvc.com
  40. [40]Microsoft-OpenAI Partnership Status 2026: What Changed ↗vaasblock.com
  41. [41]ChatGPT Statistics 2026: Users, Revenue, Which Figure to Use ↗missiongrowth.io
  42. [42]ChatGPT Pricing 2026: Free, Go, Plus and Pro Plans Compared ↗felloai.com
  43. [43]ChatGPT Plus Price by Country: 39 Markets (2026) - Code eSIM ↗codeesim.com
  44. [44]OpenAI and Deutsche Telekom partner up for new AI push ↗rcrwireless.com
  45. [45]IBM Partners with OpenAI to Accelerate Secure AI Deployment for Enterprises ↗newsroom.ibm.com
  46. [46]OpenAI reportedly plans to double its workforce to 8,000 employees - Engadget ↗engadget.com
  47. [47]Sam Altman admits OpenAI 'totally screwed up' its GPT-5 launch ↗finance.yahoo.com
  48. [48]Sam Altman Interview 2026: OpenAI CEO on AGI, Startups and the Future (beststartup.us) ↗beststartup.us
  49. [49]Sam Altman Says the World Is Going to Have to Suffer for the Glories of AI ↗futurism.com
  50. [50]Sam Altman Claims AI Singularity Arrived [2026] ↗tech-insider.org
This brief was fully generated by artificial intelligence from publicly available sources. Verify key facts before relying on it.
Prepared with BRIF · BRIF

Get a brief like this for your next guest.

Type the name, get the brief in 15 minutes: what they have been asked, what to ask instead, every claim linked to its source. The first one is free.

Get my free brief